With breakthroughs in data processing and pattern recognition through deep learning technologies,the use of advanced algorithmic models for analyzing and interpreting soil spectral information has provided an efficien...With breakthroughs in data processing and pattern recognition through deep learning technologies,the use of advanced algorithmic models for analyzing and interpreting soil spectral information has provided an efficient and economical method for soil quality assessment.However,traditional single-output networks exhibit limitations in the prediction process,particularly in their inability to fully utilize the correlations among various elements.As a result,single-output networks tend to be optimized for a single task,neglecting the interrelationships among different soil elements,which limits prediction accuracy and model generalizability.To overcome this limitation,in this study,a multi-task learning architecture with a progressive extraction network was implemented for the simultaneous prediction of multiple indicators in soil,including nitrogen(N),organic carbon(OC),calcium carbonate(CaCO 3),cation exchange capacity(CEC),and pH.Furthermore,while incorporating the Pearson correlation coefficient,convolutional neural networks,long short-term memory networks and attention mechanisms were combined to extract local abstract features from the original spectra,thereby further improving the model.This architecture is referred to as the Relevance-sharing Progressive Layered Extraction Network.The model employs an adaptive joint loss optimization method to update the weights of individual task losses in the multi-task learning training process.展开更多
Based on the classic filter of progressive triangulated irregular network（TIN） densification, an improved filter is proposed in this paper. In this method, we divide ground points into grids with certain size and se...Based on the classic filter of progressive triangulated irregular network（TIN） densification, an improved filter is proposed in this paper. In this method, we divide ground points into grids with certain size and select the lowest points in the grids to reconstruct TIN in the process of iteration. Compared with the classic filter of progressive TIN densification（PTD）, the improved method can filter out attached objects, avoid the interference of low objects and obtain relatively smooth bare-earth. In addition, this proposed filter can reduce memory requirements and be more efficient in processing huge data volume. The experimental results show that the filtering accuracy and efficiency of this method is higher than that of the PTD method.展开更多
In recent years,the convolutional neural networks(CNNs)for single image super-resolution(SISR)are becoming more and more complex,and it is more challenging to improve the SISR performance.In contrast,the reference ima...In recent years,the convolutional neural networks(CNNs)for single image super-resolution(SISR)are becoming more and more complex,and it is more challenging to improve the SISR performance.In contrast,the reference image guided super-resolution(RefSR)is an effective strategy to boost the SR(super-resolution)performance.In RefSR,the introduced high-resolution(HR)references can facilitate the high-frequency residual prediction process.According to the best of our knowledge,the existing CNN-based RefSR methods treat the features from the references and the low-resolution(LR)input equally by simply concatenating them together.However,the HR references and the LR inputs contribute differently to the final SR results.Therefore,we propose a progressive channel attention network(PCANet)for RefSR.There are two technical contributions in this paper.First,we propose a novel channel attention module(CAM),which estimates the channel weighting parameter by weightedly averaging the spatial features instead of using global averaging.Second,considering that the residual prediction process can be improved when the LR input is enriched with more details,we perform super-resolution progressively,which can take advantage of the reference images in multi-scales.Extensive quantitative and qualitative evaluations on three benchmark datasets,which represent three typical scenarios for RefSR,demonstrate that our method is superior to the state-of-the-art SISR and RefSR methods in terms of PSNR(Peak Signal-to-Noise Ratio)and SSIM(Structural Similarity).展开更多
In current small batch and customized production mode,the products change rapidly and the personal demand increases sharply.Human-robot cooperation combining the advantages of human and robot is an effective way to so...In current small batch and customized production mode,the products change rapidly and the personal demand increases sharply.Human-robot cooperation combining the advantages of human and robot is an effective way to solve the complex assembly.However,the poor reusability of historical assembly knowledge reduces the adaptability of assembly system to different tasks.For cross-domain strategy transfer,we propose a human-robot cooperative assembly(HRCA)framework which consists of three main modules:expression of HRCA strategy,transferring of HRCA strategy,and adaptive planning of motion path.Based on the analysis of subject capability and component properties,the HRCA strategy suitable for specific tasks is designed.Then the reinforcement learning is established to optimize the parameters of target encoder for feature extraction.After classification and segmentation,the actor-critic model is built to realize the adaptive path planning with progressive neural network.Finally,the proposed framework is verified to adapt to the multi-variety environment,for example,power lithium batteries.展开更多
The increasing trend towards independent fruit packaging demands a high appearance quality of individually packed fruits.In this paper,we propose an improved YOLOv5-based model,YOLO-Banana,to effectively grade banana ...The increasing trend towards independent fruit packaging demands a high appearance quality of individually packed fruits.In this paper,we propose an improved YOLOv5-based model,YOLO-Banana,to effectively grade banana appearance quality based on the number of banana defect points.Due to the minor and dense defects on the surface of bananas,existing detection algorithms have poor detection results and high missing rates.To address this,we propose a densitybased spatial clustering of applications with noise(DBSCAN)and K-means fusion clustering method that utilizes refined anchor points to obtain better initial anchor values,thereby enhancing the network’s recognition accuracy.Moreover,the optimized progressive aggregated network(PANet)enables better multi-level feature fusion.Additionally,the non-maximum suppression function is replaced with a weighted non-maximum suppression(weighted NMS)function based on distance intersection over union(DIoU).Experimental results show that the model’s accuracy is improved by 2.3%compared to the original YOLOv5 network model,thereby effectively grading the banana appearance quality.展开更多
Multienergy loads in integrated energy sys-tems(IESs)exhibit strong volatility and randomness,and existing multitask sharing methods often encounter nega-tive migration and seesaw problems when addressing complexity a...Multienergy loads in integrated energy sys-tems(IESs)exhibit strong volatility and randomness,and existing multitask sharing methods often encounter nega-tive migration and seesaw problems when addressing complexity and competition among loads.In line with these considerations,a short-term multienergy load joint prediction method based on seasonal-trend decomposition using LOESS(STL)and convolutional progressive lay-ered extraction(CPLE)is proposed,called STL-CPLE.First,STL is applied to model regular and uncertain load information into interpretable trend,seasonal,and re-sidual components.Then,joint modeling is performed for the same type of components of multienergy loads.A one-dimensional convolutional neural network(1DCNN)is constructed to extract deeper feature information.This approach works in concert with the progressive layered extraction sharing method,and convolutional shared and task-specific experts are developed to acquire common and distinctive representations of multienergy loads, re-spectively. Task-specific parameters are gradually sepa-rated through progressive routing. Finally, a subtask network is built to learn temporal dependencies using long short-term memory (LSTM). Simulation validation is performed on the IES dataset at the Tempe campus of Arizona State University, and the experiments show that the STL-CPLE method exhibits higher prediction accu-racy than do the other methods.展开更多
文摘With breakthroughs in data processing and pattern recognition through deep learning technologies,the use of advanced algorithmic models for analyzing and interpreting soil spectral information has provided an efficient and economical method for soil quality assessment.However,traditional single-output networks exhibit limitations in the prediction process,particularly in their inability to fully utilize the correlations among various elements.As a result,single-output networks tend to be optimized for a single task,neglecting the interrelationships among different soil elements,which limits prediction accuracy and model generalizability.To overcome this limitation,in this study,a multi-task learning architecture with a progressive extraction network was implemented for the simultaneous prediction of multiple indicators in soil,including nitrogen(N),organic carbon(OC),calcium carbonate(CaCO 3),cation exchange capacity(CEC),and pH.Furthermore,while incorporating the Pearson correlation coefficient,convolutional neural networks,long short-term memory networks and attention mechanisms were combined to extract local abstract features from the original spectra,thereby further improving the model.This architecture is referred to as the Relevance-sharing Progressive Layered Extraction Network.The model employs an adaptive joint loss optimization method to update the weights of individual task losses in the multi-task learning training process.
基金Supported by the National Natural Science Foundation of China(41301519)
文摘Based on the classic filter of progressive triangulated irregular network（TIN） densification, an improved filter is proposed in this paper. In this method, we divide ground points into grids with certain size and select the lowest points in the grids to reconstruct TIN in the process of iteration. Compared with the classic filter of progressive TIN densification（PTD）, the improved method can filter out attached objects, avoid the interference of low objects and obtain relatively smooth bare-earth. In addition, this proposed filter can reduce memory requirements and be more efficient in processing huge data volume. The experimental results show that the filtering accuracy and efficiency of this method is higher than that of the PTD method.
基金This work was supported in part by the National Natural Science Foundation of China under Grant Nos.61672378,61771339,and 61520106002.
文摘In recent years,the convolutional neural networks(CNNs)for single image super-resolution(SISR)are becoming more and more complex,and it is more challenging to improve the SISR performance.In contrast,the reference image guided super-resolution(RefSR)is an effective strategy to boost the SR(super-resolution)performance.In RefSR,the introduced high-resolution(HR)references can facilitate the high-frequency residual prediction process.According to the best of our knowledge,the existing CNN-based RefSR methods treat the features from the references and the low-resolution(LR)input equally by simply concatenating them together.However,the HR references and the LR inputs contribute differently to the final SR results.Therefore,we propose a progressive channel attention network(PCANet)for RefSR.There are two technical contributions in this paper.First,we propose a novel channel attention module(CAM),which estimates the channel weighting parameter by weightedly averaging the spatial features instead of using global averaging.Second,considering that the residual prediction process can be improved when the LR input is enriched with more details,we perform super-resolution progressively,which can take advantage of the reference images in multi-scales.Extensive quantitative and qualitative evaluations on three benchmark datasets,which represent three typical scenarios for RefSR,demonstrate that our method is superior to the state-of-the-art SISR and RefSR methods in terms of PSNR(Peak Signal-to-Noise Ratio)and SSIM(Structural Similarity).
基金the National Key Research and Development Program of China(No.2019YFB1706300)the National Natural Science Foundation of China(No.52075094)。
文摘In current small batch and customized production mode,the products change rapidly and the personal demand increases sharply.Human-robot cooperation combining the advantages of human and robot is an effective way to solve the complex assembly.However,the poor reusability of historical assembly knowledge reduces the adaptability of assembly system to different tasks.For cross-domain strategy transfer,we propose a human-robot cooperative assembly(HRCA)framework which consists of three main modules:expression of HRCA strategy,transferring of HRCA strategy,and adaptive planning of motion path.Based on the analysis of subject capability and component properties,the HRCA strategy suitable for specific tasks is designed.Then the reinforcement learning is established to optimize the parameters of target encoder for feature extraction.After classification and segmentation,the actor-critic model is built to realize the adaptive path planning with progressive neural network.Finally,the proposed framework is verified to adapt to the multi-variety environment,for example,power lithium batteries.
基金supported by the Beijing Science Foundation(No.9232005)the Beijing Municipal Philosophy and Social Science Foundation of China(No.19GLB036)the Beijing Science and Technology Project(No.Z221100005822014)。
文摘The increasing trend towards independent fruit packaging demands a high appearance quality of individually packed fruits.In this paper,we propose an improved YOLOv5-based model,YOLO-Banana,to effectively grade banana appearance quality based on the number of banana defect points.Due to the minor and dense defects on the surface of bananas,existing detection algorithms have poor detection results and high missing rates.To address this,we propose a densitybased spatial clustering of applications with noise(DBSCAN)and K-means fusion clustering method that utilizes refined anchor points to obtain better initial anchor values,thereby enhancing the network’s recognition accuracy.Moreover,the optimized progressive aggregated network(PANet)enables better multi-level feature fusion.Additionally,the non-maximum suppression function is replaced with a weighted non-maximum suppression(weighted NMS)function based on distance intersection over union(DIoU).Experimental results show that the model’s accuracy is improved by 2.3%compared to the original YOLOv5 network model,thereby effectively grading the banana appearance quality.
基金supported by the National Natural Sci-ence Foundation of China Joint Fund Program(No.U22A20224).
文摘Multienergy loads in integrated energy sys-tems(IESs)exhibit strong volatility and randomness,and existing multitask sharing methods often encounter nega-tive migration and seesaw problems when addressing complexity and competition among loads.In line with these considerations,a short-term multienergy load joint prediction method based on seasonal-trend decomposition using LOESS(STL)and convolutional progressive lay-ered extraction(CPLE)is proposed,called STL-CPLE.First,STL is applied to model regular and uncertain load information into interpretable trend,seasonal,and re-sidual components.Then,joint modeling is performed for the same type of components of multienergy loads.A one-dimensional convolutional neural network(1DCNN)is constructed to extract deeper feature information.This approach works in concert with the progressive layered extraction sharing method,and convolutional shared and task-specific experts are developed to acquire common and distinctive representations of multienergy loads, re-spectively. Task-specific parameters are gradually sepa-rated through progressive routing. Finally, a subtask network is built to learn temporal dependencies using long short-term memory (LSTM). Simulation validation is performed on the IES dataset at the Tempe campus of Arizona State University, and the experiments show that the STL-CPLE method exhibits higher prediction accu-racy than do the other methods.
